19 hours ago
- Fear of AI disruption spreads in concentric circles, but employees of frontier AI labs are the least apprehensive due to their wealth and dynamic labor market.
- Countries without frontier AI face the biggest risks to economic welfare and physical security, as they may be permanently pushed to the periphery.
- Even if AI is safe and aligned, it can still disrupt countries that do not build frontier AI themselves, threatening their institutions and global relevance.
- The U.S. has domestic backstops (political power, labor market flexibility) that protect it from a permanent underclass, but these protections do not extend to other nations.
- Export controls and access restrictions on frontier AI models (e.g., Anthropic, OpenAI) create a two-month head start for U.S. firms, compounding advantages over time.
- The fast-follower business model is unstable due to rising compute costs and the shift away from open-source models, making frontier dominance more likely.
- Zero-sum dynamics in economic and strategic competition mean that countries without frontier AI capture risks (e.g., crime, economic disruption) but not benefits.
- Middle powers are tempted by self-sufficiency and protectionism, but this leads to stagnation and relative decline as the frontier AI economy accelerates.
- Protectionist policies create path dependencies that isolate economies and make reentry painful, as seen in historical examples like Latin America in the 1970s.
- The alternative is to lean into disruption: secure access to frontier AI through infrastructure deals, use labor market flexibility (e.g., flexicurity), and focus on economic bottlenecks like semiconductors, manufacturing, or healthcare.
- Countries that avoid the permanent periphery must bear short-term costs for long-term gains, fighting for a place in the post-AI global economy.